1909 points · schappim · comments

# Summary The HackerNews community largely celebrates Tim Cook's transition to Executive Chairman and John Ternus's appointment as CEO, praising Cook's financial stewardship (Apple grew from $350B to $4T market cap) while expressing hope that Ternus's hardware engineering background will revitalize Apple's software quality and innovation culture. Many commenters acknowledge Cook's operational excellence but feel Apple has become risk-averse and services-focused, viewing Ternus as a potential catalyst for a new era of bold product development.

1324 points · ramonga · comments

# Summary The discussion reflects divided opinions: supporters praise replaceable batteries as reducing e-waste and planned obsolescence while enabling longer device lifespans, while skeptics worry the regulation will compromise phone design (thickness, waterproofing), increase costs, and solve a problem most consumers don't prioritize—and that software updates and security patches, not batteries, are the real drivers of phone replacement.

662 points · meetpateltech · comments

# Summary Kimi K2.6, a Chinese open-source coding model, shows impressive benchmark results claiming to match or exceed Opus 4.6, though real-world performance reports are mixed with some users finding it comparable to Sonnet while others note limitations on specialized tasks. The discussion highlights both excitement about competitive open-source alternatives to frontier US models and skepticism about whether benchmarks translate to practical use.

637 points · mfiguiere · comments

# Summary The discussion reveals skepticism about Qwen3.6-Max's claims, with commenters noting it's compared against outdated Claude versions, concerns about China's shift toward closed-source and expensive models, and a recurring theme that real-world performance often differs from benchmarks—with many users finding cheaper alternatives like GLM 5.1 or local open-weight models equally or more effective for their needs.

588 points · HotGarbage · comments

# Summary The Onion's licensing deal to take over Infowars remains pending court approval, with commenters debating the merits of the satirical takeover—some praising it as poetic justice against Alex Jones and supporting Sandy Hook families, while others question whether the reputational and financial investment is worthwhile or worry about free speech implications.

575 points · kevcampb · comments

# Summary Atlassian is automatically enrolling all users in AI training data collection for their products (Jira, Confluence, Bitbucket), with the opt-out setting reportedly broken or missing, prompting widespread criticism and speculation about a potential Anthropic acquisition alongside frustration over the company's deteriorating product quality.

365 points · speckx · comments

# Summary The discussion reveals deep skepticism about AI resistance efforts, with commenters arguing that data poisoning is ineffective against model training, while others defend the right to protest AI's real economic and societal impacts. Overall, there's disagreement about whether AI resistance is genuinely growing or merely visible in echo chambers, with most technical commenters dismissing sabotage tactics as futile.

164 points · GreenGames · comments

The discussion challenges inflated performance claims in a demo repo, arguing the 207 tok/s headline uses speculative decoding (which trades quality for speed) and cherry-picks metrics, while the actual practical throughput is roughly half that; broader critiques question the value of optimizing small models and highlight the need for better open-source inference tools and non-CUDA hardware support.

132 points · startages · comments

# Summary A blog post about AI chatbots' web scraping patterns was criticized for unclear writing (possibly AI-generated itself) and methodological issues, though readers found the core observation interesting—that different AI models fetch URLs with different priorities and patterns, reflecting their distinct architectures rather than uniform "scraping" behavior.

36 points · JumpCrisscross · comments

# Summary Tesla owners have mixed experiences with Full Self-Driving: some report impressive performance with minimal interventions, while others feel misled by unfulfilled promises, leading to a class-action lawsuit affecting thousands of California owners (though many are blocked by arbitration agreements).

44 points · EGreg · comments

The paper claims extreme KV cache compression, but commenters argue it's theoretically flawed: you can trivially achieve "infinite compression" by recomputing KV vectors on-the-fly, making the claimed gains meaningless without demonstrating practical speedups. The work appears to be purely theoretical with no empirical validation from a first-time author outside ML.

70 points · andsoitis · comments

# Summary The discussion reflects deep ambivalence about tech companies like Palantir, Anduril, and SpaceX modernizing U.S. military capabilities: while acknowledging their technological advantages over legacy defense contractors, commenters express serious concerns about the concentration of power in the hands of controversial figures, erosion of civil liberties, and potential conflicts of interest.

21 points · SilverElfin · comments

# Summary The discussion dismisses the rumor as likely fake (3-day-old Reddit account) but debates whether an acquisition would make strategic sense—some see value in Atlassian's enterprise customer base and software development data for training Claude, while others argue Anthropic could build similar tools from scratch and that Atlassian's legacy codebase is more liability than asset.

136 points · alegd · comments

# Summary The consensus is to **focus on one side of the marketplace first** by either being that side yourself (founders as drivers/couriers) or heavily subsidizing/incentivizing them, then manually match supply and demand in a narrow geographic area until you prove the model works before scaling. However, multiple comments warn this specific business idea has serious legal and safety risks around drug trafficking and customs violations.

5 points · Brajeshwar · comments

No comments available.

5 points · teleforce · comments

# Summary Amazon is investing up to $25B in AI startup Anthropic and committing $100B to cloud infrastructure, raising questions about whether this represents a genuine partnership or Amazon's attempt to gain control over a competitive AI company.

312 points · speckx · comments

179 points · akashwadhwani35 · comments

268 points · anjel · comments

424 points · kyriakosel · comments

112 points · clumsysmurf · comments

428 points · thomasp85 · comments

318 points · doener · comments

52 points · Vortigaunt · comments

48 points · Cider9986 · comments

55 points · cybermango · comments

63 points · speckx · comments

311 points · mikhael · comments

35 points · kiyanwang · comments

38 points · akyuu · comments

29 points · jaredwiener · comments

350 points · FiddlerClamp · comments

27 points · toomanyrichies · comments

24 points · Teever · comments

37 points · speckx · comments

44 points · schappim · comments

24 points · pkaeding · comments

65 points · tasoeur · comments

24 points · bizgrayson · comments

194 points · bishwasbh · comments

31 points · geox · comments

22 points · MaximilianEmel · comments

15 points · petethomas · comments

13 points · rawgabbit · comments

135 points · adunk · comments

133 points · llmmadness · comments

96 points · slicktux · comments

14 points · lxm · comments

82 points · speckx · comments

103 points · zorrn · comments